ifandelse.com | Author of machina & postal | Powerfully dumb | Planning next moves...
Joined July 2009
- Tweets 12,857
- Following 788
- Followers 1,683
- Likes 2,883
436 Photos and videos
Pinned Tweet
Apr 21
Available for hire. 🤓
Not going to list ten buzzwords or claim I'm passionate about synergy.
I ❤️ building things & have a long history of shipping working software = building teams.
node/web/full stack. Remote or hybrid. DMs are open.
1
1
4
293
May 20
If you've been putting off trying Hermes - just do it.
I'm having more fun than I should be legally allowed to...
1
49
May 16
Hermes Agent folks - I'd love to hear what models you're using (for both main and aux tasks) 🤓
54
Jim Cowart retweeted
May 15
I strongly believe there are entire companies right now under heavy AI psychosis and its impossible to have rational conversations about it with them. I can't name any specific people because they include personal friends I deeply respect, but I worry about how this plays out.
I lived through the great MTBF vs MTTR (mean-time-between-failure vs. mean-time-to-recovery) reckoning of infrastructure during the transition to cloud and cloud automation. All those arguments are rearing their ugly heads again but now its... the whole software development industry (maybe the whole world, really).
It's frightening, because the psychosis folks operate under an almost absolute "MTTR is all you need" mentality: "its fine to ship bugs because the agents will fix them so quickly and at a scale humans can't do!" We learned in infrastructure that MTTR is great but you can't yeet resilient systems entirely.
The main issue is I don't even know how to bring this up to people I know personally, because bringing this topic up leads to immediately dismissals like "no no, it has full test coverage" or "bug reports are going down" or something, which just don't paint the whole picture.
We already learned this lesson once in infrastructure: you can automate yourself into a very resilient catastrophe machine. Systems can appear healthy by local metrics while globally becoming incomprehensible. Bug reports can go down while latent risk explodes. Test coverage can rise while semantic understanding falls. Changes happens so fast that nobody notices the underlying architecture decaying.
I worry.
512
1,903
15,330
1,587,853
Jim Cowart retweeted
May 14
There's no way we get out of 2026 without OpenAI making similar moves to Anthro today. You simply cannot have extremely limited compute effectively unlimited demand and expect tokens to stay subsidized, no matter how valuable it is to have the affinity of the once subsidized.
9
1
25
4,686
May 8
Uh-oh.
Next stop: the gravity well of XML beckons.
Lots of Java guys I know who can't wait to flex those XML/XSD muscles again. 😱
HTML is the new markdown.
I've stopped writing markdown files for almost everything and switched to using Claude Code to generate HTML for me. This is why.
1
61
May 8
Just learning about the tough news coming out of Cloudflare today. The folks I follow from there have been some of the most amazing ppl to watch and learn from. If you're one of them and I can ever help throw leads your way, DMs are open.
96
Jim Cowart retweeted
Jan 14
don’t fall for the trap of waterfall development - our industry learned this lesson two decades ago. you can’t know and plan everything upfront - neither can agents.
9
31
335
102,223
May 6
So good 🎯
And everything you don't weed out that should be is there for your agents to assume as canonical representation of what you want.
78
May 6
Ughh - meanwhile those of us looking for work are fighting this signal-to-noise ratio.
May 6
Our team ran an audit on a senior engineering role last week.
80% of the applicants were fraudulent 😬
50% were stolen-identity. The other 50% were fully invented personas.
The auto-apply-at-scale stuff I get. That has a self-interest even if it's naive..
But the fraudulent stuff? What do you gain from that? Even if you somehow miraculously land the offer, you're not surviving a week of real engineering. Is the play to bluff your way through the human interview, land a remote gig, and try to be overemployed? I genuinely don't know yet.
What are others seeing?
85
May 6
I hope it means less of these, b/c man...all I've been seeing this AM
Effective today, we are:
1) Doubling Claude Code’s 5-hour rate limits for Pro, Max, and Team plans;
2) Removing the peak hours limit reduction on Claude Code for Pro and Max plans; and
3) Substantially raising our API rate limits for Opus models.
58
May 5
👀
Have been wondering if/when sub-quadratic leap would happen.
Can't wait to check this out...
May 5
Introducing SubQ - a major breakthrough in LLM intelligence.
It is the first model built on a fully sub-quadratic sparse-attention architecture (SSA),
And the first frontier model with a 12 million token context window which is:
- 52x faster than FlashAttention at 1MM tokens
- Less than 5% the cost of Opus
Transformer-based LLMs waste compute by processing every possible relationship between words (standard attention).
Only a small fraction actually matter.
@subquadratic finds and focuses only on the ones that do.
That's nearly 1,000x less compute and a new way for LLMs to scale.
1
338
May 5
Knowing what to build is still the key problem.
We've built entire process methodologies to ease and control this. All of them share the non-negotiable of time.
Agents can help grease the wheels, but knowing what you really need still takes time and iteration.
33
May 4
Tempted to ping that old colleague and tell him all the frontier models agree it's pronounced "jason", not "jay-SAHN".
54
Jim Cowart retweeted
May 4
You know another name for Agents? Workers.
7
4
116
10,731
May 4
AI has not changed human nature. You still need to sleep and touch grass.
A 2003 study showed that after 10 days at 6/hrs sleep per night, you're cognitive impairment is equal to 24-48 hours of no sleep. The decline continues with no bottom-plateau found during the study.
2
52
May 4
Don't give into FOMA (Fear of Missing Agents). Your agents are a you-amplifier. If you're burning the candle at both ends, you're not thinking clearly. Your bad decisions will be amplified.
1
32
May 4
Agents won't make up for that cognitive impairment. You have a living, breathing world model, abductive reasoning, causal understanding, metacognition, real-world-semantic-groundedness, intent, & a few-shot learning ability - all things LLMs lack.
Amplify your best-rested self.
30